bio2rdf / bio2rdf-scripts

Scripts that Bio2RDF users have created to generate RDF versions of scientific datasets
http://bio2rdf.org/
Other
129 stars 46 forks source link

KEGG Target should "Link" to individual target entities #432

Closed maulikkamdar closed 8 years ago

maulikkamdar commented 8 years ago

Hi,

I encountered the following triples indicating the Targets (CYP3A) of Cobicistat:

kegg:D09881 kegg_vocabulary:target kegg_resource:52292de0a46622a5ed543f7ab93393f8.
kegg_resource:52292de0a46622a5ed543f7ab93393f8 kegg_vocabulary:link kegg:HSA_1551 1576 1577 64816.

However, the link in the second triple is not traversable to retrieve information on the corresponding set of enzymes (which are different - CYP3A4, CYP3A5 etc.).

I think the object in the second triple should be split into the corresponding enzyme URIs?

maulikkamdar commented 8 years ago

This actually creates a problem for SPARQL queries (at least using the Triple Pattern Fragments Server) because <http://bio2rdf.org/kegg:HSA_146 147 148> is not considered a valid URI.

On another note, I am surprised it did not raise a warning when I converted to HDT, because there was another case of the following triple <http://bio2rdf.org/kegg_vocabulary:tRNA T01001> <http://purl.org/dc/terms/identifier> "kegg_vocabulary:tRNA T01001" where it shot an exception for the subject URI.

micheldumontier commented 8 years ago

ok. should work now. thanks for reporting all the errors!